Now that the therapeutic and preventive benefits of hormone replacement therapy (HRT) for the climacteric years and postmenopause have been well established,1,2 a new challenge is to enhance patient compliance. An acceptable vaginal bleeding pattern during HRT might contribute to that goal. Combined regimens of estrogen and progestin are advocated for women with an intact uterus to ensure adequate protection for the endometrium.3,4 In a sequential combined HRT regimen, progestin is added to estrogen for 10–14 days per cycle resulting in regular withdrawal bleeding at the end of the progestin phase in the majority of women. To avoid periodic withdrawal bleeding, a regimen consisting of continuous administration of estrogen and progestin has been proposed. This is expected to induce endometrial inactivity and amenorrhea.5 Unfortunately, most data on continuous combined HRT use show high rates of breakthrough bleeding,6 and research efforts have so far failed to resolve questions regarding the specific role of steroid hormones in the occurrence, duration, and severity of endometrial bleeding during HRT. The purpose of the present study was to investigate whether there is a relationship between serum estradiol (E2) or serum progestin concentrations and the occurrence of breakthrough bleeding.
Materials and Methods
The study was conducted at the Menopause Research Centre of the Free University Hospital, Amsterdam, The Netherlands. The study protocol was approved by the University Review Board and informed consent was obtained from all participants. The design of this randomized, double-blind study was aimed primarily at establishing the optimal dosage of dydrogesterone for protection of the endometrium when given in a continuous combined regimen with E2. The eligibility of the participants was determined by a period of amenorrhea of at least 12 months before baseline screening associated with serum E2 (less than 40 pg/mL) and FSH (greater than 35 mIU/mL) levels in the menopausal range. In addition, a baseline endometrial biopsy had to show either atrophy or inactivity, and women had to be free of any contraindication to estrogen use and not to be taking medications known to influence the study parameters. All participants were required to discontinue HRT 3 months prior to the first screening visit.
During the prestudy evaluation, information was obtained on medical history and number of cigarettes smoked per day, and all subjects underwent a physical and pelvic examination and a mammogram. Subjects were instructed to take their study medication in the morning and to record tablet intake in their diary on a daily basis, together with the presence of unusual signs and symptoms, the use of concomitant medication, and the occurrence of vaginal bleeding.
Subjects who satisfied all criteria were assigned randomly to one of four treatment arms: all received 2 mg micronized 17β-E2 (Zumenon; Solvay Pharmaceuticals, Weesp, The Netherlands) continuously combined with one of the four doses (2.5 mg, 5 mg, 10 mg, or 15 mg) of dydrogesterone (Duphaston; Solvay Pharmaceuticals) in an oral formulation for 168 days (six cycles of 28 days).
Dydrogesterone is a retroprogesterone with a molecular structure closely related to natural progesterone. Like most other C-21 steroids, dydrogesterone has a high affinity for progesterone receptors, a low antigonadotrophic activity, and a well-documented antiestrogenic activity, but no estrogenic or androgenic activity. In sequential combined regimens the effect of 10 mg of dydrogesterone is comparable with 10 mg medroxyprogesterone acetate. Ten milligrams of dydrogesterone represents the optimum dose of dydrogesterone for endometrial safety in a sequential combined HRT regimen.7,8 Randomization for this continuous combined HRT study was carried out (permuted blocks) in the ratio of 2:2:2:1 for the four dydrogesterone doses (2.5 mg, 5 mg, 10 mg, and 15 mg). The intentionally chosen difference in sample size of groups I (2.5 mg), II (5 mg), and III (10 mg) compared with group IV (15 mg) was related to the assumption that the minimum effective dose of dydrogesterone for endometrial protection when applied in a continuous regimen with 2 mg 17β-E2 would be 10 mg or less because a previous study with 2 mg 17β-E2 in a sequential combined regimen with 10 mg dydrogesterone indicated that the 10-mg dydrogesterone dose fulfilled the criterion of sufficient endometrial protection and that a higher dydrogesterone dose did not further increase the protective effect.
Medication was separately packaged for each participant, labeled with a code number in a double-blind manner, and dispensed by the University Hospital Pharmacy.
Subjects were examined again after 84 days and at the end of the study (day 168) or at the end of their participation. An endometrial biopsy was obtained by suction curettage (Vabra; Berkeley Medevices Inc., Berkeley, CA) without dilatation of the cervix or anesthesia at baseline and at the end of study. All tissue samples were placed immediately into formol-acetic acid for histologic assessment. All assessments were performed independently by two pathologists. Subjects were classified as current smokers (all regular cigarette smokers) or nonsmokers, according to their smoking habits during the trial. All subjects, including those who did not complete the study, underwent a final interview for counseling about further treatment and appropriate follow-up.
The pharmacokinetics of estrogens are complicated;9 with a 2-mg 17β-E2 once-daily oral dosing regimen, the plasma elimination half-life is approximately 17 hours, and steady state plasma estradiol levels are achieved after 10–21 days (Solvay Pharma Report H.102.6004.1992). Following oral administration, dydrogesterone is absorbed rapidly and metabolized to (20-α) dihydrodydrogesterone, the main metabolite. Dihydrodydrogesterone is the major source of progestational activity in plasma. Plasma elimination half-life values for dihydrodydrogesterone vary between 14 and 17 hours, and steady state plasma conditions are achieved within 3 days of dydrogesterone administration. A time interval of at least 17 hours between tablet intake and collection of blood samples was considered appropriate for assessment of serum E2 and dihydrodydrogesterone trough levels. As concomitant intake of food may affect the extent of E2 absorption, venous blood samples for serum E2 and dihydrodydrogesterone levels, routine hematology, FSH levels, and biochemistry at baseline, day 84, and day 168 were collected after an overnight fast of at least 10 hours. All determinations were carried out after centrifugation at 3300 rpm for 15 minutes at room temperature. Follicle-stimulating hormone was measured with an immunometric assay (Luminescence; Amerlite, Amersham, UK) (intra-assay and interassay coefficients of variation for FSH values are 5% and 6% at a concentration of 25 and 35 mIU/mL, respectively). Serum E2 concentrations were determined using the Baxter-Dade direct iodine-125 radioimmunoassay (E2 sensitivity, 20 pg/mL; interassay coefficient of variation for E2, 16%; intra-assay coefficient of variation for E2, 11%). All assays were performed in duplicate, and individual assays were monitored by quality control samples provided with each kit. A possible cross reactivity with estrogen conjugates10 was tested and could not be demonstrated. All assays were performed in duplicate. Serum dihydrodydrogesterone concentration was determined in duplicate using a validated gas chromotography method with electron capture detection.
The reference period method for menstrual record data analysis as proposed by the World Health Organization11 was used as a matrix for the description and analysis of all bleeding data. A minimum of three cycles of 28 days is recommended as reference time for analysis of bleeding patterns to collect sufficient events within the period for interpretation and for detection of trends in bleeding patterns over time. Assessment of the occurrence of breakthrough bleeding in relation to serum E2 and dihydrodydrogesterone levels was carried out in the second phase of the trial (day 85–168). This avoided, as far as possible, the considerable intraindividual variability in bleeding that occurs at the start of any continuous combined HRT regimen,6 while leaving sufficient observation time for a proper assessment of bleeding patterns. Data from bleeds originating in the last treatment week (day 161–168) were excluded because an endometrial biopsy, which often induces some bleeding, was carried out during this period. Bleeding episodes originating in the first phase but extending into the second phase of the trial were included in the analysis of the first phase. Bleeding episodes that started during the second phase but extended after the end of this phase (after day 168) were included in the calculation of the number of bleeding episodes but excluded from calculation of the mean duration of each episode. “No bleeding” was defined as no days of bleeding entered throughout the reference period. As exposure time to study medication differed between women who did and did not complete the study, bleeding events in the two groups could not be compared per se. To overcome this problem, spells of bleeding were determined; these spells were calculated by dividing the number of actual bleeding episodes (multiplied by 100) by the number of days of actual drug exposure.
Statistical analysis was performed using the Statistical Package for Social Sciences (SPSS/PC+4.0; SPSS Inc., Chicago, IL). The major analytic tool used to assess the relationship between serum hormone levels and the occurrence of bleeding was logistic regression analysis. The dependent variable was the occurrence of bleeding, and either serum estradiol levels or serum dihydrodydrogesterone levels served as the independent variable. Cigarette smoking also was entered as an independent variable because it showed a statistically significant interaction with serum E2 levels in a multiple regression model. No significant interactions were found for other possible independent variables including age, length of amenorrhea, and body mass index (BMI). Baseline characteristics of the study population, divided according to eligibility for analysis (evaluable sample, nonevaluable sample, and dropouts), were assessed using one-way analysis of variance for age and BMI, Kruskal-Wallis test for postmenopausal age, and χ2 for smoker status. Differences in the number of bleeding episodes and bleeding days per bleeding episode between the groups were assessed by the Kruskal-Wallis test. In the reference period, intraindividual changes in serum E2 concentrations were assessed with paired t test, and intraindividual changes in serum dihydrodydrogesterone concentrations were assessed using the Wilcoxon signed rank test. Because no significant changes were observed, the average serum E2 and dihydrodydrogesterone values at day 84 and day 168 were calculated and entered as a continuous variable in the regression analysis. Logistic regression analysis was used to estimate the occurrence of breakthrough bleeding (yes/no) in relation with various serum E2 cut-off levels. Statistical significance was set at the .05 level of probability.
One hundred ninety-four healthy postmenopausal women with an intact uterus (mean age, 51 years; range 30–61 years) and without contraindication to hormone therapy were recruited into the study. The study was completed by 177 women (91% of baseline population). Dropouts were due to unacceptable vaginal bleeding (six), persistent headache (two), weight gain (one), breast tenderness (one), mood changes (one), and no improvement of climacteric complaints (one). In addition, three women were withdrawn by the investigator due to noncompliance, and two declined further assessment. This group of 17 women were included in the dropout population in all analyses. After limiting the study group further to women who also had complete data on bleeding during the reference period and adequate blood samples at day 84 and day 168, 152 women (78% of baseline population, 86% of women who completed the study) remained evaluable (evaluable population). Table 1 summarizes the prestudy characteristics of the study population (194 women) divided into three groups: the evaluable population (152 women), the nonevaluable population (25 women), and dropouts (17 women). There were no statistically significant differences between groups. The allocation of dydrogesterone in the evaluable sample was: 2.5 mg (n = 41), 5 mg (n = 45), 10 mg (n = 43), and 15 mg (n = 23) in accordance with the randomization procedure.
Table 2 summarizes the bleeding pattern characteristics during the course of the study. There were no statistically significant differences between groups with respect to the number of subjects with amenorrhea throughout the study or for the number of spells. The number of bleeding days per bleeding episode ranged from 1 to 49 days and the median from 4 to 6 days. The number of bleeding days per event was highest in the dropout group, and this differed significantly from the evaluable and nonevaluable populations. Women in the dropout population did not have more spells than women in the other populations, but the duration of each bleeding episode was significantly longer. Vaginal bleeding was recorded in 106 of 172 (62%) subjects in the reference period. When compared with the first part of the study (day 1–84), there was no significant decrease in the number of bleeding episodes or increase in amenorrhea. As expected, serum hormone (E2 and dihydrodydrogesterone) concentrations on day 84 and day 168 showed a wide interindividual range (E2, 20–280 pg/mL; dihydrodydrogesterone, 0.51–31.85 ng/mL), but intraindividual hormone concentrations assessed on day 84 and 168 did not differ significantly (E2, P = .374; dihydrodydrogesterone, P = .255). This showed that individual serum E2 and dihydrodydrogesterone through concentrations were at steady state during the reference period. The average serum concentration of E2 and dihydrodydrogesterone for each subject was calculated by dividing the sum of the hormone concentrations on day 84 and day 168 by two.
Table 3 shows that, in multiple regression analysis with serum E2 concentration as the dependent variable, only current cigarette smoking had a statistically significant (negative) impact on serum E2 concentrations; a significant impact could not be demonstrated for the other independent variables of age, length of amenorrhea, and BMI. Cigarette smoking was entered as a binary variable (yes/no) in the logistic regression. The average serum E2 and dihydrodydrogesterone concentrations during the reference period of each subject were entered as a continuous variable. Age, BMI, and period of amenorrhea also were considered in the model (not shown). Overall, serum E2 concentrations were statistically significant related to the occurrence of breakthrough bleeding (odds ratio, 1.019; 95 % confidence interval [CI] 1.006, 1.033; P = .003). Such a relationship could not be demonstrated for serum dihydrodydrogesterone concentrations nor for current smoking habits (Table 4). The statistical power does not seem impressive, but when bleeding and nonbleeding are plotted against serum E2 levels (continuous variable) and dihydrodydrogesterone levels (divided in quartiles according to oral dose, as single dose kinetics are linear in the range 2.5 mg to 20 mg dydrogesterone) (Figure 1), it shows that bleeding occurs more above serum E2 levels of 40 pg/mL. The odds ratios for the occurrence of bleeding episodes are 2.656 (95% CI 1.454, 5.609) for serum E2 levels greater than 40 pg/mL and 2.9929 (95 % CI 1.4808, 6.0437) for serum E2 levels greater than 50 pg/mL.
There was no evidence of endometrial disease. All patients had an atrophic or inactive endometrium at baseline. At the end of the study, 117 of 152 patients still showed atrophy or inactivity of the endometrium; a single mitotic figure was demonstrated in the specimens of the remaining 35 women. Neither hyperplasia nor neoplasia was found.
This prospective study investigated whether the occurrence of breakthrough bleeding during a continuous combined HRT regimen is related to serum E2 or serum progestogen (ie, dydrogesterone) concentrations. Based on the findings from 152 postmenopausal women, we found that the serum E2 concentration is related to the occurrence of bleeds, irrespective of serum dihydrodydrogesterone concentration. The risk for the occurrence of bleeding increased significantly at serum E2 concentrations greater than 40 pg/mL. A relationship between higher oral doses of estrogen in a continuous combined HRT regimen and the occurrence of breakthrough bleeding was reported in earlier studies. Clisham et al12 and Magos et al13 suggested that higher estrogen doses (ie, 1.25 mg compared with 0.625 mg conjugated equine estrogen) were associated with lower rates of amenorrhea. Similarly, in a review of data on bleeding patterns in women who used continuous combined 2 mg 17β-estradiol, 1 mg estriol, and norethisterone acetate for up to 9 years, Dören and Schneider14 found significantly higher serum estradiol concentrations in the 13 women who experienced bleeding than in the 57 women without bleeding on the same regimen. Our findings are in accordance with these observations. Following oral administration, micronized estradiol is absorbed readily but metabolized extensively into mainly estrone (E1)-conjugates; E1-conjugates are responsible for an overestimation of serum E2 concentrations in time-resolved fluoroimmunoassays.10 All serum E2 samples were double-checked for a possible crossreactivity with estrogen conjugates; none was demonstrated.
Current cigarette smoking showed a significant negative impact on serum E2 concentrations. This antiestrogenic effect of cigarette smoking on serum E2 concentrations in postmenopausal women receiving orally administered estrogen has been reported by other authors.15,16 The mechanisms involved still are not understood precisely, but a stimulation of the metabolism of estrogen induced by polycyclic hydrocarbons has been suggested.17
As an independent variable, together with serum E2 and dihydrodydrogesterone concentrations, cigarette smoking alone did not significantly contribute to the occurrence of bleeding.
Various attempts to reduce the bleeding rates on continuous combined HRT regimens by varying the progestogen dose have been made with conflicting results. Some authors reported lower rates of uterine bleeding with higher oral doses of progestogens,13,18,19 whereas others found no difference in the incidence of amenorrhea with varying oral doses of progestogens.20,21 In the current study, serum dihydrodydrogesterone concentrations had no significant influence on the occurrence of bleeding.
The processes of endometrial bleeding are precisely controlled by a number of systemic and local factors and are centered around the function and integrity of endometrial blood vessels. Attention has focused recently on changes in the endometrial microvasculature following abnormal patterns of sex steroid exposure.22 However, the specific role of estrogen and progestin (alone or in conjunction) in this process remains to be fully elucidated, notwithstanding that withdrawal of steroid hormones is known normally to lead to withdrawal bleeding.23 A direct effect of estradiol on blood vessels leading to bleeding seems unlikely because of the temporal gap between maximal estrogen levels and the start of menstruation seen during fertile life. Furthermore, continuous unopposed estrogen administration in the postmenopause does not inevitably result in bleeding. An estrogen- or estrogen receptor–mediated effect could be involved.
Studies with sequential combined HRT regimens of 17β-E2 and dydrogesterone show that dydrogesterone protects the postmenopausal endometrium from estrogenic overstimulation24,25 and has a predominant role in providing good cycle/bleeding control.26 A dydrogesterone dosage effect on the occurrence of bleeding was not demonstrated in these studies. When 17β-E2 and dydrogesterone are combined in a continuous regimen, as in the current study, dydrogesterone continues to provide adequate endometrial protection, but its influence on bleeding control again could not be demonstrated as bleeding episodes occur irrespective of the prevailing dihydrodydrogesterone concentrations. This might indicate that the effect of dydrogesterone on glandular cells, and thus on the histology of the endometrium, correlates poorly with its effects on the endometrial processes involved in endometrial bleeding.
The results from this study point in the direction of the prevailing estrogen concentrations as a strong actor in the occurrence of bleeding episodes when combined with dydrogesterone in a continuous regimen. This could have important clinical implications. Because this association has not been reported before, confirmation is needed, preferably in studies with HRT combinations with other progestogens, before it can be used as a management tool in clinical practice.
1. te Velde ER, Leusden van HAIM. Hormonal treatment for the climacteric: Alleviation of symptoms and prevention of postmenopausal disease. Lancet 1994;343:654–8.
2. Grady D, Rubin SM, Petitti DB, Fox CS, Black D, Ettinger VL, et al. Hormone therapy to prevent disease and prolong life in postmenopausal women. Ann Intern Med 1992;117:1016–37.
3. Grady D, Gebretsadik T, Kerlikowske K, Ernster V, Petitti D. Hormone replacement therapy and endometrial cancer risk: A meta-analysis. Obstet Gynecol 1995;85:304–13.
4. Woodruff JD, Pickar JH. Incidence of endometrial hyperplasia in postmenopausal women taking conjugated estrogens (Premarin) with medroxyprogesterone acetate or conjugated estrogens alone. Am J Obstet Gynecol 1994;170:1213–23.
5. Mattson LA, Cullberg G, Samsioe G. Evaluation of a continuous oestrogenprogestogen regimen for climacteric complaints. Maturitas 1982:4;95–102.
6. Udoff L, Langenberg P, Adashi EY. Combined continuous hormone replacement therapy: A critical review. Obstet Gynecol 1995;86:306–16.
7. Ferenczy A, Gelfand MM. Endometrial histology and bleeding patterns in postmenopausal women taking sequential, combined estradiol and dydrogesterone. Maturitas 1997;26:219–26.
8. Burch DJ, Spowart KJM, Jesinger DK, Randall S, Smith SK. A dose-ranging study of the use of cyclical dydrogesterone with continuous 17 β oestradiol. Br J Obstet Gynaecol 1995;102:243–8.
9. Stumpf PG. Pharmacokinetics of estrogen. Obstet Gynecol 1990;75:9S–14S.
10. Thomas CMG, Van den Berg RJ, Segers MFG, Bartelink ML, Thien T. Inaccurate measurements of 17β-oestradiol in serum of female volunteers after oral administration of milligram amounts of micronised 17β-oestradiol. Clin Chem 1993;39:2341–2.
11. Belsey EM, Farley TMM. The analysis of menstrual bleeding patterns: A review. Contraception 1988;38:129–56.
12. Clisham PR, de Ziegler D, Lozano K, Judd HJ. Comparison of continuous versus sequential estrogen and progestin therapy in postmenopausal women. Obstet Gynecol 1991;77:241–6.
13. Magos AL, Brincat M, Studd JWW, Wardle P, Schlesinger P, O'Dowd O. Amenorrhoea and endometrial atrophy with continuous oral estrogens and progestogen therapy in postmenopausal women. Obstet Gynecol 1985;65:496–9.
14. Dören M, Schneider HPG. Long-term compliance of continuous combined estrogen and progestogen replacement in postmenopausal women. Maturitas 1996;25:99–105.
15. Baron JA, La Vecchia C, Levi F. The antiestrogenic effect of cigarette smoking in women. Am J Obstet Gynecol 1990;162:502–14.
16. Cassidenti DL, Vijod AG, Vijod MA, Stanczyk FZ, Lobo RA. Short-term effects of smoking on the pharmacokinetic profiles of micronized estradiol in postmenopausal women. Am J Obstet Gynecol 1990;163:1953–60.
17. Michnovicz JJ, Hershcopf RJ, Naganuma H, Bradlow HL, Fishman J. Increased 2-hydroxylation of estradiol as a possible mechanism for the anti-estrogenic effect of cigarette smoking. N Engl J Med 1986;315:1305–9.
18. Archer DF, Pickar JH, Bottiglioni F, for the Menopause Study Group. Bleeding patterns in postmenopausal women taking continuous combined or sequential regimens of conjugated estrogens with medroxyprogesterone acetate. Obstet Gynecol 1994;83:686–92.
19. Sporrong T, Hellgren M, Samsioe G, Mattson LA. Comparison of four continously administered progestogen plus estradiol combinations for climacteric complaints. Br J Obstet Gynaecol 1988;95:1042–8.
20. Hillard TC, Siddle NC, Whitehead MI, Fraser DI, Pryse-Davies J. Continuous combined conjugated equine estrogen-progestogen therapy: Effects of medroxyprogesterone acetate and norethin-drone acetate on bleeding patterns and endometrial histologic diagnosis. Am J Obstet Gynecol 1992;167:1–7.
21. Weinstein L. Efficacy of a continuous estrogen-progestin regimen in the menopausal patient. Obstet Gynecol 1987;69:929–32.
22. Hickey M, Lau TM, Russell P, Fraser LS, Rogers PAW. Micro vascular density in conditions of endometrial atrophy. Hum Reprod 1996;11:2009–13.
23. Tabibzadeh S. The signals and molecular pathways involved in human menstruation, a unique process of tissue destruction and remodeling. Mol Hum Reprod 1996;2:77–92.
24. Ferenczy A, Gelfand MM. Endometrial histology and bleeding patterns in postmenopausal women taking sequential, combined estradiol and dydrogesterone. Maturitas 1997;26:219–26.
25. Gelfand MM, Fugere P, Bissonnette F, Wiita B, Yang HM, Lorrain J, et al. Conjugated estrogens combined with sequential dydrogesterone or medroxyprogesterone acetate in postmenopausal women: Effects on lipoproteins, glucose tolerance, endometrial histology and bleeding. Menopause 1997;4:10–18.
26. van de Weijer PHM, Mol M, Biesheuvel E. Vaginal bleeding patterns during sequential regimens of oestradiol and dydrogesterone. Eur Menopause J 1995;4:9–16.